Metadatanormalisering

Metadatanormalisering is the process of converting metadata from disparate sources into a consistent, interoperable form. It involves applying standardized formats, controlled vocabularies, and crosswalks so that metadata from different systems can be understood and combined.

The goals include improving search and discovery, enabling data integration, ensuring quality and consistency, and reducing redundancy. Normalized metadata supports reliable filtering, ranking, and interoperability across platforms.

Techniques include standardizing value formats (dates to ISO 8601, language codes to ISO 639-1/3), normalizing geographic identifiers to standard URIs, and resolving or canonicalizing identifiers (e.g., DOIs, ORCID iDs). It also includes mapping between schemas (MARC, Dublin Core, MODS, schema.org) and applying controlled vocabularies for subjects, genres, and roles.

Practical aspects involve validation, using crosswalks and metadata registries, and enrichment from external authorities. Tools include metadata transformers, validation schemas, and cataloging guidelines.

Challenges include heterogeneous data quality, legacy records, evolving vocabularies, and the balance between normalization and preserving domain-specific nuance. Over-normalization can erase useful local practices; thus, processes often include provenance and versioning.

Applications are common in libraries, archives, museums, digital repositories, and large-scale data platforms that require integrated search and data exchange.